What if your glasses could remember the doctor’s advice you forgot two days later?
That is the pitch behind Mira AI glasses, a wearable that transcribes conversations, translates speech in real time, and stores searchable summaries of what you’ve said and heard. On paper, it sounds like a memory upgrade. In daily life, it looks a lot more complicated.
Peter O’Brien tested Mira’s glasses for FRANCE 24 and came away with two clear impressions: the tech can feel magical, and wearing it around other people can feel awkward fast.
What Mira’s smart glasses are trying to do
Mira’s pitch is simple and unsettling at the same time. The glasses sit on your face and act like a running memory log. They can transcribe what people say, show captions in your field of view, and translate spoken language on the fly. According to the video’s description, they can live-caption 60 languages.
Peter demonstrated how that works in conversation. If someone switched from English to Chinese or Arabic, the glasses could place translated text in front of his eyes in little green letters. A small ring controller lets the wearer switch options without reaching for a phone.
The glasses also plug into AI tools that go beyond note-taking. Peter said he could use them to answer questions in real time or fact-check a claim mid-conversation. If someone said Apple sold a record number of phones last year, the system could check that and return an answer on the spot.
Then there is the archive. Mira connects to an iPad or iPhone and keeps a record of conversations the user chooses to transcribe. Those records are searchable, and each conversation gets an AI summary. The company calls that a “second brain.” That phrase sounds like marketing, but it also describes the promise clearly. The glasses are meant to remember what you missed, what you forgot, and what you may want later.
The startup behind the product also has a backstory that helps explain the attention around it. Mira was founded by Harvard dropouts AnhPhu Nguyen and Caine Ardayfio, the pair who previously drew notice for hacking Meta’s smart glasses to show how privacy could break down. Peter said they had raised about $6.5 million and shipped roughly 1,600 pairs, mostly to professionals with heavy meeting schedules. A TechCrunch report on Mira’s always-on AI glasses points to the same broad idea: a wearable that listens, transcribes, and tries to sit between your brain and your calendar.
“The goal for this is to become like the closest thing to Jarvis that you can have.”
That line captures both the appeal and the risk. Mira is selling memory, speed, and AI help, all in a device that lives inches from your eyes.
Where the glasses felt useful in real life
The strongest part of Peter O’Brien’s test was not the demo. It was the ordinary stuff.
He wore the glasses for a couple of weeks, and one moment stood out. During a doctor’s visit, he forgot, or tuned out, a piece of advice. A few days later, he spotted that advice in the AI summary. That turned the product from a flashy gadget into something more practical. The glasses had remembered what he had chosen not to.
That use case makes immediate sense. People forget details from medical appointments, work meetings, and quick hallway chats all the time. A searchable summary can feel less like a toy and more like a safety net.
Another moment came while Peter was recovering with a broken leg. He lay on the sofa in one room and used the glasses to tell his computer in another room what to do. He found that impressive, although he also said it was slow. That detail matters. Many wearable demos sound smooth, but speed decides whether a feature becomes habit or novelty.
The live translation also worked well in his testing. He described it as fast. Still, he questioned how often he would use it in his own life. That is the recurring theme with AI wearables right now. A feature can be smart, polished, and real, yet still have a narrow place in someone’s day.
Peter had a less flattering reaction to the fact-checking tool. He joked that using it made him sound even more insufferable. That line landed because it cuts through the product pitch. A device that can correct people in real time may be technically impressive, but social settings are not product demos. No one wants lunch with a person whose glasses are always waiting to challenge the table.
So the real test is not whether Mira can do these things. It can. The harder question is whether those abilities fit the messy rhythm of normal life.
The price, features, and the bigger race around AI wearables
The product details show where Mira sits in the market.
Here is a quick snapshot of the features Peter highlighted:
| Category | What Peter reported | What it means | | | | | | Price | $650 upfront | This is a premium device, not an impulse buy | | Subscription | $20 per month for full AI features | The best tools cost extra over time | | Core tools | Transcription, AI summaries, translation, fact-checking | Mira is built around conversation and memory | | Controls | In-lens display plus a ring controller | It tries to stay hands-free | | Companion devices | Syncs with iPhone and iPad | Your data lives beyond the glasses | | Camera | No camera | Mira avoids one of the biggest smart-glasses privacy fears | | Bonus use | Built-in auto-cue | It can double as a teleprompter |
The price puts Mira in serious-tech territory. At $650, plus a monthly fee, these glasses are aimed at people who can make a case for them at work or who simply like testing expensive new hardware.
Peter also noted that Mira is not alone. Meta sells Ray-Ban smart glasses that cost more, and in his view they do less than Mira’s glasses. He said Meta had already bought one of Mira’s competitors and had been linked to plans for an AI pendant. Amazon, he added, had acquired another competitor in the same space.
That matters because it shows this is not one odd product chasing headlines. Large tech firms are spending money around the idea that AI should move off the phone and onto the body. Glasses, pins, pendants, and other wearables all aim at the same basic goal: capture what you say, understand what you mean, and help before you ask.
A separate profile of Mira’s founders and early funding paints the same picture of a startup trying to move fast while larger companies circle the category.
Still, the table above also reveals the catch. Mira offers plenty of AI tools, but none of them erase the cost, the monthly bill, or the social weight of putting a conversation machine on your face.
Why a phone still makes more sense most of the time
The sharpest question in the segment came from the host: why not use your phone?
It is an obvious question because most of what Mira does already exists in some form on a smartphone. Phones can record audio, transcribe meetings, translate languages, summarize notes, and answer questions with AI. The glasses promise speed and convenience, but the phone already sits in most people’s pocket.
Peter did not dodge that point. He said that during the moments when the glasses felt useful, he often had them resting on a table or hanging around his collar. In other words, he did not always need them on his face. He also said you do not need the display for many of the AI features.
That gets to the heart of the product problem. The closer Mira moves to “phone, but wearable,” the more it has to justify why the wearable form matters. If the same job gets done with a device people already accept, the glasses need to offer a clear gain.
Instead, Mira often creates a trade-off. You gain hands-free access, but you lose social ease. Phones are familiar. No one stops a meeting because you placed your phone on the table. Glasses with visible tech inside them are different. Peter said these looked a bit odd, and from some angles you could see the display. Then there is the ring controller. Together, they signal that something unusual is happening.
That visual strangeness creates friction before the product even does anything. People notice. They wonder if they are being recorded or analyzed. The user then has to explain the device, the transcription, and the terms of consent. A phone rarely asks that much of a room.
So the phone comparison is not only about features. It is about what society already tolerates. Mira may win on novelty and occasional convenience. The phone still wins on comfort, habit, and trust.
Privacy is the part Mira cannot smooth over
Peter kept returning to the same issue: telling people they were being transcribed felt uneasy.
That reaction is hard to dismiss because it gets to the real barrier for products like this. The hardware may work. The AI may work. The social contract is still shaky.
To Mira’s credit, the glasses do not include a camera. That choice matters because cameras were central to the earlier privacy fears around hacked Meta smart glasses. By leaving cameras out, Mira removes one obvious problem. Yet transcription alone is enough to trigger concern.
Peter pointed to France as one example. There, recording someone without permission is against the law. He also questioned whether Mira’s method, discarding audio and keeping only a transcript, changes that enough to settle the issue. He did not claim a legal answer, and that uncertainty is part of the story.
The concern grows once you consider what the product stores. One of the founders said that 3 million words from his conversations had already been transcribed and were searchable in the app. Searchable is the key word there. Search turns talk into data. It lets a machine map your habits, your work, your worries, and your relationships.
The founder described the result in striking terms. He said the glasses knew his situation better than almost anyone in his life, including family members who speak with him all the time. That statement is revealing because it shows what this device becomes if you use it often. It is not a notebook. It is a private record of your spoken life.
That may sound useful when you are tracking meetings or recalling advice. It sounds less comfortable when you picture every casual remark passing into a searchable archive. Memory has always faded at the edges. Mira tries to stop that fade. Privacy often depends on it.
Peter’s final judgment was plain: until passive transcription becomes more normal in society, wearing something like this will stay awkward. The problem is not the quality of the transcript. The problem is that other people have to live inside your memory system too.
Where Mira fits today, and where it still falls short
Peter did not dismiss the product. He also did not pretend it was ready for broad daily use.
He suggested there are settings where Mira makes more sense right now. Work meetings are one. In that context, people already expect notes, summaries, and follow-up records. A wearable that handles those tasks could fit, especially if everyone knows it is in use. Journalists and reporters could also find value in hands-free note capture, especially when pulling out a phone would be clumsy or distracting.
Then there is the most amusing feature in the segment, the one Peter used the whole time: auto-cue. The glasses can show text in front of the wearer, turning them into a hidden teleprompter. He admitted, on air, that he had been reading from the glasses through the report. The joke worked because it showed a real benefit. In television, that kind of discreet prompt is handy.
At the same time, that moment also exposed what wearables often do best right now. They do not replace everything. They slot into narrow tasks and perform them well enough to earn a place with certain users.
That is why Mira feels more like a specialist tool than a mass-market device. If your day is full of interviews, meetings, languages, scripts, or dictated tasks, the glasses may have a purpose. If your day looks more ordinary, the product has to fight harder for wrist time, pocket space, and trust.
The company wants to build a second brain. Peter’s test suggests it has built the outline of one. The sharper memory is real. The translation is real. The teleprompter trick is real. So is the discomfort that follows them around.
Final thoughts
Mira’s glasses show that AI wearables can be useful before they become normal. That is a real step forward, especially when a device helps you recover forgotten details or work hands-free.
But the strongest lesson from Peter O’Brien’s test is not about software. It is about consent. Until people are comfortable being transcribed by the person sitting across from them, smart glasses like Mira will stay most appealing in controlled settings, not everyday life.
A second brain sounds helpful. Wearing one in public still sounds like a negotiation.
